One of the unpleasant things about cancer is its tenacity. It hangs on, to reappear years after it should have been removed and the patient presumably cured. Even in 1878, when A. Von Winiwarter summarized the results of the German surgeons, it was realized that patients had to be observed for several years before conclusion could be reached that a successful procedure had been carried out. An arbitrary 3-year period was accepted as a minimum during which patients should be followed, and this was soon increased to 5 years because experience showed that 3 years was inadequate to conclude about the eventual fate of the patients. For some cancers even this period is not long enough, and an indefinite length of time for periodic examinations is now recommended. Physicians are increasingly more reluctant to speak of “5-year cure,” and prefer the longer but more noncommittal and realistic expression of “5-year survival without clinical evidence of disease.”
What has been this experience? What has been achieved with the use of surgery and of radiation in the treatment of cancer?
Our most satisfactory information on these points comes from the registry of all cancer patients in the State of Connecticut, which has been maintained since 1935, and the more recent collection and analysis of records on cancer patients under the end results evaluation program of the National Cancer Institute. This program gathers records from 100 selected hospitals in the United States, and also compares similar statistics from several Scandinavian countries and from England.
There is no question but that overall we are doing a bit better in the salvage of patients with cancer, year by year. In 1935, in Connecticut, approximately 25 percent of all patients with cancer were alive 5 years later. Now the proportion is over 30 percent. Thus, one of every three patients with cancer lives 5 years or longer following the diagnosis.
Analysis of the records shows that this improvement cannot be attributed to better treatment alone, but also reflects better awareness of the disease by the patients. This is particularly true for cancer of the uterine cervix and for the cancers of the mouth. Also, the increased survival of cancer patients is due to better results in the treatment of a few finite sites of cancer: this is especially true of cancer of the rectum and large intestine, in both men and women, cancer of the body of the uterus as well as of the cervix, cancer of the larynx, and of the thyroid. The results in cancer of the esophagus, of the stomach, of the lung, and of the brain, unfortunately remain dismally stable and dismally low. Figure 4 summarizes some of these findings and trends, as presented at the Fourth National Cancer Conference in 1960.
It has been stated that only for cancer of the skin can we be reasonably satisfied with our results. It has also been stated that the end results in cancer could be further improved, perhaps to the level of 50 percent survival at 5 years, if our presently available knowledge were truly applied. This application involves the earliest reference of people to medical attention, and the earliest recognition and effective treatment of cancer by the physician.
Skeptical but thoughtful critics of these statistics point out that the true effects of treatment can be measured only if we know how similar untreated patients would behave. What proportion of the 5-year survivals, they ask, are really attributable to the natural course of the disease?
This is an important question, the answer to which influences not only the selection of treatment, but our concepts of cancer as a disease process. Is cancer, or at least are some cancers, really not localized at all, but merely one manifestation of generalized disease? The sum of evidence indicates that most cancers do start as a localized process, but it is also true that some of them may become widespread very early.
It is medically and ethically impossible to maintain individuals with cancer as “untreated controls,” as is often demanded in a laboratory study with animals. The few instances of people who refuse treatment for religious and other reasons are seldom representative of the whole population of cancer patients. Nevertheless, there are available some older observations, particularly from England and from Massachusetts, that demonstrate the inexorable course of cancer in man. The course can be slow, especially in certain types of cancer. Cancer of the breast is the best example, for there almost 20 percent of the women who have no definite treatment are alive 5 years later; of course, they are alive with actively growing cancers. In cancer of the rectum, of the urinary bladder, and of the uterus, 5 to 10 percent of patients survive for 5 years or longer without definite treatment. In any compilation of 5-year end-results, therefore, a proportion that may be as high as 10 percent may be due to the natural duration of this group of chronic diseases rather than to the beneficial effects of treatment.
We also know that rarely there occurs a remarkable “spontaneous” recovery from cancer. By the time cancer is clinically diagnosed, such events have been estimated to happen only once in perhaps 100,000 patients. For practical reasons, there-fore, recovery from cancer without surgical or radiation treatment is too rare to have an influence. But the theoretical implications of such observations are obviously significant. If we understood better the circumstances of these rare cancer regressions, we would be in a position to attempt to produce them under experimental conditions and, eventually, in patients.
In cancer, which is fatal unless its course is interrupted by definitive treatment, the most obvious first question in the evaluation of any new treatment is whether it improves the survival of the patients. If the new procedure is not lifesaving, it may still be useful in producing temporary arrest of the disease, or clinical improvement of the patient by removing distressing symptoms. And even increased survival should be achieved at a reasonable price of risk and morbidity, so that the life that is saved is worth living. How are we to obtain the necessary answers to these complex questions in the setting of clinical medicine? These same questions also apply to the comparative value of old, sometimes well accepted, methods in the treatment of cancer.
We know that the comparisons of the results of several hospitals or of different surgeons do not give entirely satisfactory answers. This is because patients that enter different hospitals may be quite different in many ways, and because different surgeons may select patients for treatment by using quite different criteria. The greatest single factor that makes cancer clinically curable is its extent. If one surgeon includes in his groups of patients a considerable proportion with clinical spread, his results will be much lower than those of a surgeon who rigorously excludes such patients from his treatment. Some very useful comparisons are still possible if the clinical and other “stages” of the patients entering different treatment groups are carefully recorded and are taken into consideration in the analyses.
A new phase of scientific medicine can be predicted through the acceptance of the techniques of designed clinical trials for obtaining answers to many vexing problems regarding cancer treatment. For this advance we owe a debt to our British colleagues, for their pioneer role in the development of clinical trials.
A number of designed clinical trials in cancer have been performed in England and Denmark, as well as in the United States. In Denmark, two operations for breast cancer, the extended radical and a simpler procedure, supplemented by radiation, were compared. The results were almost identical. In the United States, several large groups of cooperating surgeons used nitrogen mustard and similar agents as additions to surgical resection for cancer of the stomach, lung, rectum, and breast. At the completion of the operation, the patients received either the drug or contents of identical containers without the drug. Under the conditions of these carefully designed trials, no benefit was shown by the additive treatment except in patients with breast cancer. In the absence of the investigations, argument may have continued for many years regarding the possible merit of such additive therapy. The clinical trials substituted numerical knowledge for subjective opinion, and provided the answers in the minimum time on a minimum number of patients.